Vesk Lab

Your video ads, read second by second.

Vesk Lab reads your Meta and TikTok video ads second by second, ranks what it finds, and tests the strongest idea against a holdout. Every claim carries a seal that says how far you can trust it, and “no difference” counts as a result.

Three of your ads read, free, when we launch.

Launch date to be announced. For US brands and agencies running paid video. See a sample report (PDF).

Readout
    Example: a 7-second stock clip from Pexels, not a client’s ad. Face and bottle tracked by software; readouts written by hand.

    Method

    Read, rank, then test.

    1. 01

      Read

      Every ad, second by second: when the face turns, when the product and the brand show up, what text is on screen, where the cuts fall, what is said. Where your account has the data, lined up with how many viewers were still playing at each second.

    2. 02

      Rank

      What we find becomes a short list of hypotheses, strongest first. Each one carries its seal, so you can tell a pattern across ads from a single observation.

    3. 03

      Test

      The strongest idea goes against a holdout, with the plan signed before launch. If the answer is “no difference”, you get that too.

    Every claim says how far it goes.

    One question, three strengths of answer: does showing the bottle first keep more people watching?

    1. Observation

      In this ad the bottle is on screen from 0:00, and the face reaches the camera at 0:03.7. It describes one ad. It does not compare.

    2. Association

      Among comparable skincare ads, the ones that open on the product held more viewers at 0:03. Audience or budget the data doesn’t show may explain it.

    3. Test

      Against a holdout, the product-first cut moved the chosen KPI by a stated amount, with a stated range. This is the only seal that means “because”.

    Example. In a study, the numbers come from your account and your test.

    Sample

    See a full report before you write to us.

    Seven pages on three public US beauty ads, for a fictional client: brands withheld, data marked as illustrative. This is what a read of three ads looks like.

    Download the sample (PDF)
    Page 1 of the sample report: summary, top hypotheses and how to read the evidence seals

    Offers

    Three ways to go further.

    1. Diagnostic

      All your live ads, read second by second and, where our base covers your niche, compared with similar public ads. You get the ranked hypotheses and the next test worth running. Any budget.

      Up to Association

    2. Audit

      Your ad account’s history, past A/B tests included, sorted into what held up and what was noise. For accounts with a history of spend.

      Up to Association

    3. Pre-registered test

      The strongest idea against a holdout, sized to your traffic, with the plan signed before launch. “No difference” is reported as a result. Media spend stays yours.

      Test
    Join the free-read list

    Scope and pricing by email, when we launch.

    Three of your ads read, free, when we launch.

    Launch date to be announced. Tell us where to find your ads, and when we open we read three of them second by second and email you what we see: when the product and the brand show up, what changes in the first two seconds, and one idea worth testing. Every line carries its seal.

    Three of your ads read, free, when we launch.

    We only write to you about Vesk Lab. Reply “stop” and we will.